Disentangling the functional relationship between neural networks of social cognition

809 

Abstract Submission 

Lara Maliske¹, Matthias Schurz², Philipp Kanske³

¹Technische Universität Dresden, Dresden, Saxony, ²Institute of Psychology and Digital Science Center, University of Innsbruck, Innsbruck, Tyrol, ³Clinical Psychology and Behavioral Neuroscience, Technische Universität Dresden, Dresden, Germany

Lara Maliske  
Technische Universität Dresden
Dresden, Saxony

Matthias Schurz  
Institute of Psychology and Digital Science Center, University of Innsbruck
Innsbruck, Tyrol

Philipp Kanske  
Clinical Psychology and Behavioral Neuroscience, Technische Universität Dresden
Dresden, Germany

Recent advances in the study of empathy and Theory of Mind (ToM) demonstrate the need to investigate the two in interaction: naturalistic settings often blur the distinction between affect and cognition and demand the simultaneous processing of such different stimulus dimensions. Here, we followed up on the results of a recent coordinate-based meta-analysis and hierarchical clustering analysis (Schurz et al., 2021), as well as a follow-up meta-analytic connectivity modeling analysis (Maliske et al., 2023) that described a hierarchical model of social affect and cognition, as well as how neural representations of empathy and ToM were enabled through network co-activation and connectivity. Specifically, we further investigated the relationship between social affective, cognitive, and complex social neural activation maps by projecting them along a principal gradient of macroscale cortical organization (Margulies et al., 2016), and comparing them to other neural activation patterns related to social cognition.

We re-analyzed data from a meta-analysis and agglomerative hierarchical clustering (Schurz et al., 2021) including 188 studies from the empathy and ToM literature across 11 more narrow tasks groups, that were clustered into three overarching clusters of neural activation related to social affect and cognition (cognitive, intermediate, affective cluster). Furthermore, we re-analyzed data from a follow-up meta-analytic connectivity modeling study (Maliske et al., 2023) including 140 studies associated with the right anterior cingulate cortex, right posterior cingulate cortex, left temporoparietal junction, and left anterior insula.
To project the meta-analytic (co-activation) maps onto a principal gradient of macroscale organization, as well as more narrow task groups, we determined overlap between neural activation maps using a variant of the dice score (indicating the percentage of voxels in i1 that overlap with i2). Specifically, we calculated the overlap between meta-analytic (co-activation) maps and a) continuous changes in the functional spectrum described by the principal gradient (in increments of 5%), as well as b) more narrow task groups related to social affect and cognition (see Schurz et al., 2021).

Meta-analytic (co-activation) maps related to social affect and cognition (affective and cognitive cluster, respectively) were positioned at separable locations along the principle gradient, while meta-analytic (co-activation) maps related to complex social cognition (intermediate cluster) tended to overlap with the locations of both the affective (located towards the sensory/ unimodal part of the gradient, 30-55th percentile) and the cognitive cluster (located towards the transmodal end of the gradient, 80-95th percentile.
In terms of overlap with specific tasks, the cognitive cluster showed most pronounced overlap with the false belief task (e.g., Saxe & Kanwisher, 2002; overlap ranging from 81-95%). Neural (co-)activation patterns related to social affect and complex social cognition (affective and intermediate cluster, respectively) showed overlap with diverse task profiles, although some trends emerged (e.g., a larger overlap of affective cluster and observing pain, as well as reading the mind in the eyes meta-analytic maps).

Re-analysis of coordinate-based meta-analyses and meta-analytic connectivity modeling allowed us to further probe the relationship between neural activation patterns related to social affect, cognition, as well as complex social tasks. More precisely, in contrasts to the notion of independence of empathy and ToM-related neural networks, the results presented here indicate that complex social tasks (intermediate cluster) rely on cross-network interaction (network integration), and their neural activation are similar to a neural activation patterns related to a range of different tasks measuring empathy and ToM.

Emotional Perception

Social Cognition ¹

Activation (eg. BOLD task-fMRI)

fMRI Connectivity and Network Modeling ²

Task-Independent and Resting-State Analysis

Meta- Analysis

Social Interactions

Other - social cognition; co-activation; functional gradient